Contrary to the general belief, rather than artificial intelligence designs and hardware deficiencies, the key to autonomous driving is the smart software that can analyze complex and changing road, traffic, and weather conditions to create and implement the most accurate action plan in a matter of seconds.
Turan noted that they propose a deep learning-based self-supervised approach for ego-motion estimation -- a robust and complementary localization solution under inclement weather conditions.
The proposed approach is a geometry-aware method that attentively fuses the rich representation capability of visual sensors and the weather-immune features provided by radars using an attention-based learning technique.
The dream of fully autonomous driving is very unlikely to become a reality in the next 10-15 years, he said, explaining: "While training the artificial intelligence algorithms for autonomous driving, it is impossible to create a data repository to cover all exceptional circumstances that a driver may encounter."
"Although technology giants such as Tesla and Google are leading the autonomous driving technology, they are still limited to relatively small-scale trials," he added.
